Spot Virtual Machines
Spot Virtual Machines (VMs) provide an affordable way to run GPU-based workloads by utilizing Exabits' unused computing capacity. This guide explains how Spot VMs work, how they differ from on-demand VMs, their limitations, and the types of use cases they are best suited for.
Overview
Spot Virtual Machines allow you to access idle GPU resources at a significantly lower cost than on-demand VMs, making them ideal for fault-tolerant workloads that do not require guaranteed uptime.
Key Differences: Spot vs On-Demand Instances
| Feature | Spot VMs | On-Demand VMs |
|---|---|---|
| Availability | Unreliable, may be terminated at any time. | Guaranteed availability during usage. |
| Data Persistence | No data persistence. | Full persistence support. |
Limitations of Spot VMs
Please note the following when using spot instances:
- No termination notice: If capacity is reclaimed, the virtual machine may be deleted at any time without prior notice.
- No data persistence: Unsaved data will be lost if the VM is terminated.
NOTE
Please note that spot instances may be terminated at any time without any notification.
Ideal use cases
- Non-critical or interruptible workloads: Tasks that can tolerate unexpected interruptions, such as background jobs, test environments, or temporary development setups.
- Scaling compute-heavy workloads: When used alongside on-demand resources, spot VMs can help increase capacity at a lower cost.